Stretching Peptides to Generate Small Molecule β-Strand Mimics

Advances in the modulation of protein–protein interactions (PPIs) enable both characterization of PPI networks that govern diseases and design of therapeutics and probes. The shallow protein surfaces that dominate PPIs are challenging to target using standard methods, and approaches for accessing extended backbone structures are limited. Here, we incorporate a rigid, linear, diyne brace between side chains at the i to i+2 positions to generate a family of low-molecular-weight, extended-backbone peptide macrocycles. NMR and density functional theory studies show that these stretched peptides adopt stable, rigid conformations in solution and can be tuned to explore extended peptide conformational space. The diyne brace is formed in excellent conversions (>95%) and amenable to high-throughput synthesis. The minimalist structure-inducing tripeptide core (<300 Da) is amenable to further synthetic elaboration. Diyne-braced inhibitors of bacterial type 1 signal peptidase demonstrate the utility of the technique.


Synthetic Methods
General Solid-Phase Peptide Synthesis (SPPS) Procedure All peptides were chain assembled on Rink Amide polystyrene resin (0.64 mmol/g) or Rink-Amide TentaGel resin (0.2 mmol/g, Rapp Polymere GmbH) by individual hand couplings. All standard amino acid couplings were carried out with the equivalent ratio of for 20 min following standard SPPS protocol with N-terminal Fmocprotection. Propargyl containing amino acids were coupled by hand using the equivalent ratio of [2.5]: [2.5]: [3.8] of these components for 90 min, and followed by a qualitative ninhydrin test to ensure complete coupling.

N-Methylation
After peptide assembly of the macrocycle ring, N-methylation of residue 3 was performed on-resin. Resin was treated twice with [1]: [1] DBU:EtTFA in DMF for 30 min, and then checked with a qualitative ninhydrin test to ensure complete protection of the N-terminus. An equivalent ratio of [5]: [10] [PPh3]: [MeOH] in DCM was added to resin in a closed tube with stopcock. 5 equivalents DIAD was added last, and the tube was vented, then shaken for 1 hr. This procedure was repeated until a test cleavage showed good conversion to the methylated N-terminus. The protecting group was cleaved with 20 equivalents of NaBH4 in 1:1 MeOH:DCM. The amino acid directly following this step was coupled twice by standard SPPS protocol.
On-Resin Glaser Coupling To a 15 mL Falcon tube was added CuCl (white powder, stored under Ar, 100 mg, 1 mmol, 10 equiv.), 2,2'-bipyridine ligand (165 mg, 1.5 mmol, 15 equiv.), and 0.1 mmol of washed Fmoc-deprotected peptideloaded resin, prior to the addition of any alkyl tails. The mixture was briefly vortexed to mix the Cu and ligand. To this tube was added 5 mL DMA, followed by 175 μL DIPEA (1 mmol, 10 equiv.). The tube was capped then parafilmed and placed horizontally in a secondary container to reduce settling of compounds to the bottom of the tube and improve interaction of the solution with O2 in the headspace. The tube was incubated at 37 C with shaking at 180 rpm for 5 hr. Following the stapling reaction, the resin was washed with DMF and 20% piperidine and subjected to either continued synthesis (i.e. coupling of alkyl tail) or standard TFA cleavage conditions for SPPS procedure.
Chloroformate Coupling Alkyl chloroformate tails were appended by adding DMAP (1 equiv.) to a mixture of resin and chloroformate (5 equiv.) in 1:1 DCM:DMF (2 mL total), stirring for 1 hr, and subsequent flow washing with DMF. Coupling completion was ensured by a qualitative ninhydrin test.
Cleavage and Purification Peptides were cleaved from resin within 48 hr of the formation of the diyne brace to prevent undesired copper-catalyzed modifications observed to occur on-resin. Peptides proceeded to standard cleavage from resin using [2.5]:[2.5]:[95] TIPS:H2O:conc. TFA at 45 C for 4 hr. The TFA was reduced via rotovap then added to an excess of 30% MeCN and lyophilized. The crude peptide thus obtained was purified via preparative reverse-phase HPLC using optimized gradients as detailed in the analytical peptide data. Safety Statement No unexpected or unusually high safety hazards were encountered.

Analytical and Purification Conditions
Crude peptides were purified on preparative HPLC (Waters Autopurify prep LC with diode array and QDa mass spec) using either the generic gradient or an optimized gradient as specified. Peptides were characterized by analytical LC/MS on a Waters Acquity I-Class UPLC with diode array and time of flight mass spec (Waters G2-XS). 5% B Purity and identity of peptides was confirmed using a Waters I-Class LC with diode array and G2-XS time of flight (TOF) mass spectrometer with the analytical gradient listed above.

Nuclear Magnetic Resonance Spectroscopy
Two main approaches were taken when preparing the braced peptide products for NMR analyses. Braced peptides not containing a lipid tail (0.8 -1.4 mg) were dissolved in 450 μL H2O and 50 μL D2O and analyzed in 5 mm tubes on a 700 MHz Bruker Avance III spectrometer equipped with a cryoprobe. Alkynomycin analogues (1.8 -2.2 mg) were dissolved in 150 μL DMSO-d6 and analyzed on a 600 MHz Bruker spectrometer equipped with either a 5 mm CPQCI or CPDCH cryoprobe in 3 mm tubes. As all compounds were purified as TFA salts, pH of solutions were adjusted to within the range 4.6 -5.3 and scans were obtained at 298 K. Water suppression was achieved by excitation sculpting during proton spectral acquisition.
1 H homonuclear data included 2D total correlation spectroscopy (TOCSY) with a mixing time of 80 ms, rotating-frame Overhouser effect spectroscopy (ROESY) with a mixing time of 100 ms, and nuclear Overhauser effect spectroscopy (NOESY) with a mixing time of 300 ms. The homonuclear data were recorded with a sweep width of 10 or 12 ppm with 4k data points in the direct and 512 increments in the indirect dimension. Heteronuclear single quantum coherence (HSQC) data were also recorded at natural abundance. The 1 H-13 C HSQC spectra were recorded with 2k data points over a sweep width of 10 or 12 ppm in the direct dimension, and 280 increments over a sweep width of 80 ppm, covering the aliphatic carbon region, in the indirect dimension. The 1 H-15 N HSQC spectra were recorded with 2k data points over a sweep width of 10 ppm in the direct dimension, and 128 increments over a sweep width of 32 ppm in the indirect dimension for the group of peptides.
All data were collected using Topspin 4.0.6 (Bruker), processed with MestReNova 14.2.1 (Maestrelab Research). Water solvated samples were referenced to internal DSS at 0.0 ppm and DMSO samples were referenced to residual solvent peak at 2.50 ppm.

S-4
Density Functional Theorem Modeling A Monte Carlo conformational search undertaken using the OPLSe3 forcefield and simulated water solvent (GB/SA) using Macromodel v12. 1 The selected conformers (<3 kcal/mol of the global minimum) were optimized by density functional theory (DFT) calculations at the B3LYP/6-31G(d,p) level with PCM implicit solvent model for water using Gaussian software G16W. 2 A single point energy of the optimized conformers was calculated using M062X/6-31+g(d,p) with PCM implicit solvent model for water and duplicate conformers and conformers with >3 kcal/mol of the global minimum were removed. Finally, single point free energy of the optimized conformers was calculated using M062X/6-31+G(d,p) with PCM implicit solvent model for water were used to scale the calculated NMR parameters relative to their Boltzmann population and the vibrational frequencies where checked for a true minimum, i.e. no negative frequencies and this energy was used to calculated the Boltzmann populations for each compound. Minimum Inhibitory Concentration Assay MICs were performed in accordance with the broth microdilution protocol from the Clinical and Standards Laboratory Institute. 3 Alkynomycins were dissolved at 10 mg/mL in DMSO and two-fold serial dilutions were made across a 96well plate containing Muller-Hinton (MH) broth. One plate was used for each indicator organism, and each well reached a final inoculum of 5×10 5 colony forming units per mL. Using OD600 readings normalized to a negative control, MICs were recorded as the lowest concentration at which growth was inhibited after a 24 hr incubation at 37 C. Controls were included on each plate to ensure peptide and media sterility. MICs are reported in Table 2 and Tables S4, S5 as an average of at least three independent trials.
Indicator organisms were stored at -80 C as glycerol stocks. Each organism was streaked onto tryptic soy agar and grown for 24 hr at 37 C. Single colonies were picked and used to inoculate MH broth. Bacteria culture was diluted to reach an OD600 of 0.015 prior to being used for MIC assays.

Molecular Docking
The co-crystallized ligand was removed from the structure (1t7d), 4 with crystallographic waters retained. The structure was prepared for docking using AutoDockTools. Grid maps were prepared using AutoGrid v4.2.6 with 30 Å in each dimension and 0.375 Å spacing. The ligand was prepared with terminal alkynes. "Glue" atoms (G0 type) appropriate for macrocyclic docking were placed 1.5 Å from the end of each of alkyne, and the terminal carbon of the alkynes were typed CG0. Docking was performed using AutoDock-GPU v1.5.3. 20 independent genetic algorithm runs were performed, with an AutoStop triggered by an energy standard deviation cutoff at 0.15 kcal/mol. The best scored pose was used for structural analysis.

S-5
Supporting Figures

S-6
For reference, Structure of Arylomycin A-C16: Note: Arylomycin refers to angles from crystal structure of Arylomycin BAL4850C. Angles are denoted as N (N-terminal residue), M (middle residue), or C (C-terminal residue).

S-53
Alkynomycin Compounds Compound 10: Alkynomycin A12 The product was purified by mass-directed prep LC. The crude material was purified on a Waters Autopurification LC with a Waters BEH C18 column (5 um, 19x160 mm) using a 0.1% aqueous trifluoroacetic acid:acetonitrile gradient (30 mL/min, main segment of gradient at 45-65% acetonitrile over 8 minutes) at ambient temperature. Fractionation was triggered by a Waters QDa single quadrupole mass spec (ESI+).

S-56
Compound S1: Alkynomycin B12 The product was purified by mass-directed prep LC.

S-68
Compound 11: Alkynomycin A16 The product was purified by mass-directed prep LC. The crude material was purified on a Waters Autopurification LC with a Waters BEH C18 column (5 um, 19x160 mm) using a 0.1% aqueous trifluoroacetic acid:acetonitrile gradient (30 mL/min, main segment of gradient at 55-75% acetonitrile over 8 minutes) at ambient temperature. Fractionation was triggered by a Waters QDa single quadrupole mass spec (ESI+).